Long-term projects with a fairly well-paved path to monetization. If industry was willing to invest in things like pure mathematics, or deep space astronomy, or just generally subsidizing smart people to investigate things they think are interesting, we wouldn’t need the NSF.

The point is that the efficiency of the private sector is a Faustian bargain: it comes at the cost of expecting an ROI for the investigators.

A small minority of the private sector is burning money on long-term investments. Another small minority are burning money on ventures which cannot be monetized. At their intersection, there exists zero companies. Yes, zero. Not close to zero, just zero. And it will always be zero.

How will you survive if what you're doing takes both a lot of time and makes no money?

Think about something like the FDA. It's a cost sink. The private sector will never do something like that because it's explicitly anti-profit.

Aren't both heavily propped up with government funding?